A

AI Agents From Scratch

Learn to build AI agents step by step without frameworks.

FrameworkOpen SourceGrowing

What is AI Agents From Scratch?

AI Agents From Scratch is learn to build AI agents step by step without frameworks.

About

AI Agents From Scratch is a hands-on learning series designed for developers who want to understand how to build AI agents from first principles. This repository provides step-by-step examples using local LLMs and node-llama-cpp, covering fundamental concepts, agent architectures, and the reasoning behind design choices in frameworks. It emphasizes learning through building and offers a progressive learning path.

Strengths

  • Hands-on approach to learning AI agent development
  • Comprehensive examples with detailed explanations
  • Focus on understanding fundamental concepts
  • No reliance on external frameworks for initial learning
  • Encourages experimentation and deep understanding

Limitations

  • Requires local setup and model downloads
  • May be complex for complete beginners
  • Limited to Node.js environment
  • Not a production-ready framework
  • Focuses on educational purposes rather than deployment

Use Cases

Building custom AI agents for specific tasksUnderstanding the inner workings of LLMsExperimenting with agent architectures and patternsLearning prompt engineering and system promptsImplementing memory and reasoning in agents

Integrations

Node.jsTypeScriptReactOpenAI APIlocal LLM models